Most computer systems include some type of non-volatile storage to store and maintain data Non-volatile storage provides a storage medium whose contents are retained if power is removed. Common examples of non-volatile storage include floppy disks, hard disks and tapes.
Most types of non-volatile storage offer relatively high reliability, particularly when redundant storage is used. The performance of non-volatile storage varies widely between different types of non-volatile storage. For example, hard disks traditionally offer higher performance, in terms of access speeds, than floppy disks or tapes. Most types of non-volatile storage can be re-used, since the process used to store data is non-destructive. When data in the form of a data file is deleted, often directory information associated with the data file is merely updated in the non-volatile storage, and the data file itself is unchanged. For example, on many computer systems, deleting a data file is accomplished by removing the filename from a file directory or file allocation table that makes the locations occupied by the data file available for other data. However, the data file still resides on the non-volatile storage and can be recovered so long as it is not overwritten with other data. This approach makes it difficult to know whether a particular copy of data is an original copy and makes the data susceptible to being obtained by a third party.
Another type of non-volatile storage allows data to be written only once, but read as many times as desired. This type of non-volatile storage is usually referred to as a write once, read many (WORM) storage medium. A common example of a WORM non-volatile storage device is an optical disk. This type of storage medium is useful for archiving data, for example certain types of medical and business records, that are written only once and may be read several times. This guarantees that a particular copy of data is an original copy since the data cannot be updated or overwritten.
Both WORM systems and conventional read/write storage systems suffer from the disadvantage that they are vulnerable to modification of data. A user of data is not sure if the data is original or modified by the unknown sources when using the data. For example, in a disk storage subsystem, an unauthorized individual may remove the disk drive and alter, intercept, or copy the information stored on the disk drive.
In addition, information stored in storage systems can become undesirable to be stored after the passage of sufficient time. Therefore, in some contexts it is desirable to provide a way for old information to expire from a storage system and become unavailable. One such example is a company record that should be destroyed after five years according to the company policy.
Based on the need to securely store and maintain data and the limitations in the prior approaches, an approach for storing and maintaining data that provides a relatively high assurance that a particular copy of data is an original copy is highly desirable.